This grant is to carry out research and software development in those areas of statistical methodology which are important to cancer clinical trials. The intended research is of a general nature and the results will have broader applicability than cancer clinical trials. The research plan has three main components which are all directly motivated and hence related to cancer clinical trials: (1) Theoretical investigation on Statistical and Computing Methodology, (2) Software Development for the Integration and Building of Systems, and (3) Combination of the Previous Two Components into Usable and Reliable Software. All software will be developed and written so that it is machine independent (Standard Fortran) and requires minimal software maintenance. It is expected that the software will be readily accessed by non-programmers (physicians, data managers) as well as statisticians.